Sam M the Demifemme

Just a demisexual demi-femme trying to be a better human and an incurable nerd.

[demisemme]

A Luta Sigue

Donate to Support Mothers Fighting Family Separation!, organized by Operations A Luta Sigue

Over the last few months, attacks against working-class immigrants… Operations A Luta Sigue needs your support for Support Mothers Fighting

A Luta Sigue is running a fundraiser to support Mothers Fighting Family Separation due to ICE raids that end in the deportation of their families' primary breadwinners

[casper-ghostly]

Anyway, daily reminder from a culturally isolated Romani person.

Gypsy does not mean wanderer.

It literally means ‘people from egypt’ or similar, as europeans believed Romani people were from Egypt. It has become known similar to nomad due to how our ancestors have been forced to be nomadic due to racism and ostracization, but it is a SLUR.

Romani people are STILL being forcibly sterilized.

Romani people are STILL being forced into ghettos.

Romani people are still facing violence and danger in countless European countries- and recently, I’ve seen the beginnings of the extremes in the United States.

Have a little fucking respect and DON’T USE A SLUR THAT’S BEEN USED FOR CENTURIES AGAINST US.

And for the love of whatever’s up there, ESPECIALLY do not use it to describe your witchcraft. It is playing on the ‘magic gypsy’ trope, and is EXTREMELY insulting.

non romani people, please reblog this.

[depsidase]

[justalurkr]

Text of tweet under the cut because it is loooong.

But... Stochastic Parrots.

Timnit Gebru was fired from Google in December 2020 for refusing to retract a research paper, and every single warning that paper made about large language models has now happened at a scale the industry spent 4 years trying to make people forget about.

Her name is Timnit Gebru.

She co-led the Ethical AI team at Google. She co-wrote a paper called "On the Dangers of Stochastic Parrots" with Emily Bender at the University of Washington and two other researchers. The paper was 14 pages long. It was submitted to a top AI ethics conference. And it was the reason Google decided that one of the most senior Black women in AI research could no longer work there.

The story Google told publicly was that she resigned. The story she told, confirmed by 2,695 of her colleagues in an open letter, was that she was fired by email while on vacation because she refused to either retract the paper or remove her name from it.

The paper had not even been published yet.

Here is what she actually wrote, and why every prediction inside it has now come true.

The first warning was about scale itself. Bender and Gebru argued that training ever-larger models on ever-larger scrapes of the internet would produce systems that appeared fluent but had no actual understanding of language. They called these systems stochastic parrots because they would repeat patterns from training data with statistical confidence and zero comprehension. The paper predicted that this apparent intelligence would fool both users and developers into trusting outputs that were structurally incapable of being reliable.

This was 2020. GPT-3 had just come out. The paper predicted the hallucination problem before anyone had a word for it.

The second warning was about bias amplification. The paper documented in detail that internet-scale training data contains systematic overrepresentation of dominant viewpoints and underrepresentation of marginalized ones. The models would not just absorb this bias. They would amplify it, because the optimization process rewards confident outputs, and confidence in language patterns tracks frequency in the training set.

The prediction was that hiring tools built on these models would discriminate against women. That healthcare triage tools would underperform on Black patients. That loan approval systems would entrench inequality while presenting their decisions as neutral algorithmic judgment.

Every one of those things has now been documented in deployment.

Amazon's hiring algorithm penalized resumes that contained the word "women" in any context. Healthcare risk scoring algorithms used by major US hospitals were found to systematically underestimate the medical needs of Black patients. Apple Card's credit algorithm gave wives credit lines 10x lower than their husbands for the same financial profile.

The third warning was about environmental cost. The paper calculated that training a single large language model produced emissions equivalent to the lifetime output of 5 cars. The prediction was that the race to scale would create an environmental footprint that would eventually rival entire industries.

In 2024, Google's emissions were up 48% from 2019, and the company explicitly blamed AI infrastructure. Microsoft's were up 29%, same reason. Both companies have now quietly abandoned the climate commitments they were publicly celebrating the year Gebru was fired.

The fourth warning was about documentation. The paper argued that the training datasets being assembled were too large for anyone to actually audit. Nobody at Google, OpenAI, Meta, or any other lab could tell you with confidence what was in the data their models were trained on. This was not a temporary problem to be solved later. It was a permanent feature of the approach.

In 2023, researchers discovered that the LAION-5B dataset, used to train Stable Diffusion and other major image models, contained thousands of images of child sexual abuse material. The companies that had trained on the dataset had no way of knowing. The paper predicted that category of failure 3 years before it was found.

The fifth warning was the one Google cared about most.

Bender and Gebru argued that the deployment of these systems would centralize linguistic and cultural power in the hands of the small number of companies that could afford to train them. The internet would become a place where the dominant voice was a statistical average of dominant voices, presented as a neutral assistant. Languages underrepresented in the training data would degrade over time as more web content was generated by these systems and fed back into the next training run.

This is now happening in real time. A 2024 study found that 57% of new web content in English is AI-generated or AI-assisted. Researchers studying low-resource languages have documented active degradation in translation quality, because the synthetic content fed back into training is itself worse in those languages.

The paper Google fired her for predicted the model collapse problem before model collapse had a name.

The mechanism behind why this all happened is the part of her work that nobody quotes.

Gebru's argument was not that AI is dangerous in some abstract sci-fi sense. Her argument was that AI is dangerous in a very specific structural sense. The technology was being built by a small group of researchers who shared similar backgrounds, worked at similar companies, and were rewarded for shipping products faster than competitors. The incentive structure made it impossible for safety, ethics, and bias concerns to slow anything down. Anyone inside the system who raised those concerns was either ignored, sidelined, or removed.

She was making that argument from inside Google.

Then Google proved her right by removing her.

The team Google had built to make sure their AI was safe was dismantled in 90 days because they did the job they had been hired to do. Margaret Mitchell, the other co-lead of the Ethical AI team, was fired two months after Gebru for searching through her own emails for evidence of how Gebru had been treated.

Gebru did not stop. She founded DAIR, the Distributed AI Research Institute, in 2021. The mission is to do AI research outside the control of the companies that have a financial interest in not hearing the answers.

Every prediction in the Stochastic Parrots paper has now been validated by deployment. Hallucinations are an industry-wide problem the largest labs cannot solve. Bias amplification has been documented in hiring, healthcare, lending, and criminal justice. Environmental costs are larger than entire small countries. Training data audits remain impossible. Model collapse is an active research crisis at every major lab.

The question worth sitting with is the one almost no one in the industry will say out loud.

Every researcher with the technical credibility to call out these problems watched what happened to her in December 2020 and made a calculation about their own career. The number of people willing to speak publicly about safety and ethics issues inside the major AI labs collapsed after that firing and has not recovered.

The researcher Google fired for warning about exactly what is now happening was right.

The company that fired her is now the second-largest deployer of the technology she warned about.

And the people inside that company who agree with her are not allowed to say so.

[irispurpurea]

This is the paper. It's excellent, highly recommend reading it.

I remember reading about Gebru's firing but I had no idea this was the paper she was fired over.

[hi-im-cy]

Seth Moulton and Jake Auchincloss are absolute garbage people. As horrible as any Republican. NEVER trust their lies.

[massachusetts-official]

Official Post of Massachusetts

[fandomshatepeopleofcolor]

If you live in Alabama, Georgia, Idaho, Kentucky, Oregon or Pennsylvania, please make sure to vote on May 19! No election or primary is too big or small, every vote matters!

[fawnpalmer]

quarterly reminder that if i reblog something ai-generated it is 110% and always an accident and for the love of god please tell me so i can delete it from my blog

[mothermara]

remember to bury the dead with a phone, everyone. these days the ferry terminal at the river styx wants you to download a fucking app

[fairycosmos]

studio ghibli movies are like if someone took all of your fondest, softest childhood dreams and put them into a film

[covid-69]

[xxwebpunkxx]

CRUCIAL NEWS FOR THOSE IN MASSACHUSETTS.

Even if you don’t live there, if you still live somewhere in the states, please, listen up.

“The MA legislature and governor just proposed dangerous social media surveillance legislation that would require everyone in the state, including adults, to upload a government ID or submit to a face scan to create a social media account. If something like this passes here, it will spread to other states faster than activists can fight back.

I have an op-ed out in today's Boston Globe explaining why this legislation would hurt kids and help Trump attack LGBTQ youth and undocumented folks. You can read it here (no paywall) https://archive.ph/rAK4P

If you're in MA, please send this to your state senators and the governor! They need to read it! You can also use this tool to easily call the key lawmakers who will be making decisions about this: https://www.stoponlineidchecks.org/ma “

- Fight For the Future Signal chat

I encourage reading EFF’s article linked below as it goes into the greater details of the legislation aside from this, but key notes:

• Forces everyone (not just kids) to upload a government ID or submit to a facial recognition scan to create a social media account or post online, even on educational sites like Wikipedia

• Force parents to send even more sensitive information to Big Tech companies to “verify” custody in order for a minor to create an account (think: sending a birth certificate to Mark Zuckerberg to let your teen create an account)

• Allow unsupportive parents to lock queer and trans youth off the Internet, restricting access to life-saving community and information with zero recourse

• Expand the Trump admin’s surveillance capabilities, making it easier for them to persecute people organizing protests, monitoring ICE, and criticizing the administration

• Cost Massachusetts tax payers enormously, as the legislation is clearly unconstitutional and would face legal challenges.

• Specifically harm undocumented folks who use social media to warn communities about ICE activity and raise money for family members

• Fine smaller and non-profit tech platforms like Mastodon and Bluesky out of business, further expanding the monopolies of Big Tech platforms

It is one of many of the worst pieces of digital surveillance legislation right now in the states.

If you’re in MA, please call and write to your state senators and governor through EFF’s article below, and if you’re not, either spread the word or call them too, but, don’t give them your ZIP code so they don’t know where you are.

Stop Online ID Checks in Massachusetts

Stop Online ID Checks in Massachusetts

[massachusetts-official]

Official Post of Massachusetts

[original-character-chaos]

Hey so if you live in massachusetts I would phone your governor right now and demand to know what the fuck she's trying to pull. If you do not live there SPREAD THIS MESSAGE.

These "protect the kids" laws are bad enough. We dont need the democrats getting ideas.

AND I MEAN NOW.

[massachusetts-official]

Official Post of Massachusetts

[ohmywyll]

CAN WE PLEASE HAVE WYLL MERCH WITHOUT MIZORA?

CAN WE PLEASE HAVE WYLL MERCH WITHOUT MIZORA?

CAN WE PLEASE HAVE WYLL MERCH WITHOUT MIZORA?

[moniquill]

[transisstor]

Reblog if you think fanfiction is a legitimate form of creative writing.

[lowkeyed1]

shoutout to the willow fans <3 3 years on from the frankly unnecessary and rude memory-holing on top of cancellation, and it's a bittersweet feeling to still be part of such a dedicated set of fans. all the creatives have moved on -- they have to, it's their job -- and we are definitely the people thinking most in the world about this special universe and everything that happened in it, everything that could have happened if we'd been allowed to go on we keep it alive love you guys <3 <3 <3

[fozmeadows]

I am so unbelievably fucking sick of AI tools. Every time I go to the doctor now, they start up by asking, "Is it all right if I use this AI voice translation software to record my notes?" while already setting it up, and then they look nonplussed when I tell them no. Like man I appreciate it takes you a modicum of extra time and concentration to type up things in my chart yourself, but that is in fact a loadbearing part of what I'm paying you to do! I want to know that the notes you take are accurate, or at least accurate to your assessment in the moment! "Oh but we read through the AI summary to check that it's accurate" I don't care! I don't fucking care!! Even if you did it right after my appointment, that's no guarantee you'll remember every beat of what was discussed, because now you're trying to remember thirty minutes' worth of conversation after the fact instead of just typing a few sentences at five minute intervals! And that's the best case scenario - how many patients do you see a day? What fucking good is it if you're reviewing all these summaries in bulk after a ten hour shift, or a day later, or two? How are you gonna catch any errors or hallucinations when you barely remember our appointment? Whereas when you type it yourself, your brain builds connections between what's being heard and what's being written. You remember more. I want you to remember more! I do not want AI in my fucking medical care! Fuck!

[fadingembers]

As a kid, I was really upset that Bill Watterson wouldn't license Calvin & Hobbes so I could have plushies or so there would be a Saturday morning cartoon. Now, I realize his resistance is the reason we don't have a Calvin & Hobbes DreamWorks movie starring Chris Pratt.